ORIGINAL INVESTIGATION

Transcription

1 ORIGINAL INVESTIGATION Subclinical Atherosclerosis and Risk of Atrial Fibrillation The Rotterdam Study Jan Heeringa, MD; Deirdre A. M. van der Kuip, MD, PhD; Albert Hofman, MD, PhD; Jan A. Kors, PhD; Frank J. A. van Rooij, MSc; Greg Y. H. Lip, MD, FRCP; Jacqueline C. M. Witteman, PhD Background: Myocardial infarction is an important risk factor for atrial fibrillation, but the role of subclinical atherosclerosis is unknown. This longitudinal study evaluates whether atherosclerosis affects the risk of atrial fibrillation in persons without overt coronary disease. Methods: This investigation was part of the Rotterdam Study, a population-based cohort study among persons 55 years or older. Participants with atrial fibrillation at baseline, with a history of myocardial infarction, or with angina pectoris and those who had undergone cardiac operative procedures were excluded, leaving 4407 subjects for the analyses. Baseline intima-media thickness of the common carotid artery and the presence of carotid plaques were used as indices of generalized atherosclerosis. During a median follow-up of 7.5 years, 269 cases of incident atrial fibrillation were identified. Relative risks were calculated with 95% confidence intervals, adjusted for age and sex, using the Cox proportional hazards model. Additional adjustments were made for body mass index, hypertension, systolic blood pressure, serum cholesterol level, smoking, diabetes mellitus, left ventricular hypertrophy on the electrocardiogram, and the use of cardiac medication. Results: The risk of atrial fibrillation was associated with carotid intima-media thickness (relative risk, 1.90; 95% confidence interval, , highest vs lowest quartile) and severity of carotid plaques (relative risk, 1.49; 95% confidence interval, , severe vs absence). Risk estimates were stronger in women than in men. Conclusions: Atherosclerosis in participants without manifest atherosclerotic disease is an independent risk factor for atrial fibrillation. These results suggest that aggressive treatment of asymptomatic atherosclerosis may help to prevent atrial fibrillation. Arch Intern Med. 2007;167: Author Affiliations: Departments of Epidemiology and Biostatistics (Drs Heeringa, van der Kuip, Hofman, and Witteman and Mr van Rooij) and Medical Informatics (Dr Kors), Erasmus Medical Center, Rotterdam, the Netherlands; and Haemostasis, Thrombosis, and Vascular Biology Unit, University Department of Medicine, City Hospital, Birmingham, England (Dr Lip). MYOCARDIAL INFARCtion is a strong predictor of atrial fibrillation, especially in men. 1 Myocardial damage, often in association with heart failure, contributes to the onset of atrial fibrillation in subjects with myocardial infarction. It is not known, however, whether the presence of coronary atherosclerosis without manifest myocardial infarction also causes atrial fibrillation. At a subclinical level, atherosclerotic vascular disease may cause some damage to myocardial tissue by gradually reducing the blood supply to the atrial tissues because of ischemia/infarction, causing premature apoptosis of myocytes, fibrous tissue replacement, and subsequent facilitation of reentry processes. 2 The role of atherosclerosis in the development of atrial fibrillation is supported by results of studies in which atrial tissues were histologically investigated. 3-5 Results of studies that examined the relation between atrial fibrillation and cardiac manifestations of atherosclerosis, such as angina pectoris in the absence of a myocardial infarction and angiographically demonstrated narrowing of coronary arteries without overt damage to the heart, are conflicting Studies dependent on coronary angiographic documentation of coronary artery disease are also limited by the selected nature of the population studied and, by the very nature of this measurement, cannot be feasibly or ethically applied to large-scale population surveys. A noninvasive index of atherosclerosis that uses ultrasonography, intimamedia thickness, and plaques has been shown to be a good indicator of systemic atherosclerotic vascular disease, with a relation to the risk of stroke and myocardial infarction. 11 To our knowledge, only 1 population-based study, the Cardiovascular Health Study, 6,7 reported on the as- 382

2 sociation between atherosclerosis and risk of atrial fibrillation. In that study, no significant associations were found between 2 measures of generalized atherosclerosis (carotid intima-media thickness and the ankle-arm index) and risk of atrial fibrillation. 6,7 In the present population-based analysis from the Rotterdam Study, we investigated the associations between intima-media thickness and plaques of the extracranial carotid arteries (as measured by ultrasonography) and risk of atrial fibrillation in asymptomatic persons during a median follow-up of 7.5 years. METHODS STUDY POPULATION The Rotterdam Study is a population-based prospective cohort study aimed at assessing the occurrence and progression of and risk factors for chronic diseases in the elderly. Neurogeriatric, cardiovascular, locomotor, and ophthalmologic diseases are the main areas of interest. 12 All residents of Ommoord, a suburb of Rotterdam, the Netherlands, who were 55 years or older were invited to participate. Of the eligible individuals, 7983 (77.7%) responded. From 1990 to 1993, participants were interviewed at their homes and 7151 were examined at the research center to obtain baseline measurements, including a 10-second, 12-lead resting electrocardiogram (ECG). Those who did not visit the research center were in general dependent or lived in nursing homes. Participants were reexamined during 2 follow-up rounds. The first round of follow-up examinations was performed between July 1, 1993, and December 31, The second round of follow-up examinations started April 1, 1997, and ended December 31, The medical ethics committee of the Erasmus University in Rotterdam approved the study, and all participants gave informed consent. MEASUREMENTS OF ATRIAL FIBRILLATION AND ATHEROSCLEROSIS Atrial fibrillation at baseline and during follow-up was ascertained using 3 methods, as described earlier. 13 In short, at baseline and during the follow-up examinations, 10-second, 12- lead ECGs were recorded with an electrocardiograph (ACTA; EsaOte, Florence, Italy), stored digitally, and analyzed by the modular ECG analysis system (MEANS) Additional information was obtained from the medical files of participating general practitioners, which included the results of their own work and those of physicians practicing in hospitals and outpatient clinics, and from a national registration of all hospital discharge diagnoses. All study participants were followed up from the date of entrance in the Rotterdam Study ( ) to the date of the onset of atrial fibrillation, death, or loss to follow-up or to January 1, 2000, whichever came first. Follow-up by January 2000 was complete for 99.1% of the study population. The common carotid artery, the bifurcation, and the internal carotid artery on both sides were visualized by means of ultrasonography using a 7.5-MHz transducer (ATL Ultramark IV; Advanced Technology Laboratories, Bethel, Wash). Common carotid intima-media thickness was measured for a 1-cm length that was proximal to the bulbus. The anterior and posterior walls in the left and right common carotid arteries were measured and averaged. The averages of the left and right side were added and divided by 2, if measurements at both sides were available. When the result of one side was absent owing to poor visualization, the result of the other side was used. A plaque was defined as a focal broadening of the intima-media relative to the adjacent segments, with protrusion into the lumen composed of only calcified deposits or a combination of calcified and noncalcified material. The left and right common carotid arteries, left and right carotid bifurcation, and left and right internal carotid arteries were examined for the presence of plaques. A weighted plaque score was obtained by counting the sites where a plaque was visible and dividing this number by the total number of sites for which an ultrasonographic image was available. The result was multiplied by 6, the maximum number of sites. This score ranges from 0 to 6. Subjects were excluded from the analyses if visualization of more than 4 of the 6 sites had not been possible. All examinations were performed with the observer blinded to the clinical details of the participants. MEASUREMENTS OF COVARIATES Information on current health status, medical history, and smoking behavior was obtained using a computerized questionnaire. Participants were classified as current smokers, former smokers, or never smokers. The body mass index was calculated as weight in kilograms divided by the square of height in meters. Blood pressure was measured twice at the right upper arm with a random-zero mercury sphygmomanometer in the sitting position. Systolic and diastolic blood pressures were calculated as the average of the 2 consecutive measurements. Hypertension was defined as a systolic blood pressure of 160 mm Hg or higher or a diastolic blood pressure of 100 mm Hg or higher or the use of blood pressure lowering drugs prescribed for hypertension. A history of myocardial infarction was defined as a self-reported myocardial infarction with hospital admission or the presence of a myocardial infarction on the ECG. A positive self-report of myocardial infarction was confirmed by reviewing the medical records of general practitioners and specialists. Angina pectoris was measured by the Rose questionnaire. 17 Left ventricular hypertrophy was diagnosed using the MEANS program with an algorithm that takes into account QRS voltages, with an age-dependent correction and repolarization. Nonfasting blood samples were drawn by means of venipuncture. Serum total cholesterol levels were measured with an automated enzymatic method. Diabetes mellitus was defined as the use of antidiabetic medication or a random or postload serum glucose level of 220 mg/dl ( 11.1 mmol/l) or higher. Cases of incident myocardial infarction were ascertained as described previously. 18 VITAL STATUS Information on vital status was obtained on a regular basis from the Central Register of Population of the municipality of Rotterdam, from collaborating general practitioners, and by obtaining information during follow-up rounds. For those participants for whom information remained missing, the Central Registry of Genealogy of the Netherlands was consulted in The Hague. This national institute receives population registry records of inhabitants of the Netherlands who have died. POPULATION FOR ANALYSIS For the present study, an ECG was not available for analysis of atrial fibrillation in 343 participants who had visited the research center at baseline, mainly owing to logistic reasons. We additionally excluded 1666 participants belonging to at least 1 of the following categories: (1) participants with a history of myocardial infarction (typical, silent, or non Q-wave) at baseline 383

3 Table 1. Baseline Characteristics of the Study Population by Sex: the Rotterdam Study, * Men Women Characteristic (n = 1669 [37.9%]) (n = 2738 [62.1%]) Age, mean ± SD, y 67.1 ± ± 8.8 BMI, mean ± SD 25.6 ± ± 4.0 Hypertension Blood pressure, mean ± SD, mm Hg Systolic ± ± 22.9 Diastolic 75.3 ± ± 11.5 Cholesterol level, mean ± SD, mg/dl Total 243 ± ± 46 HDL-C 48 ± ± 14 Smoking Current Former Never Diabetes mellitus Left ventricular hypertrophy Cardiac medication Intima-media thickness, mm Mean ± SD 0.84 ± ± 0.14 Median (interquartile 0.82 ( ) 0.78 ( ) range) None Mild Moderate Severe Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by the square of height in meters); HDL-C, high-density lipoprotein cholesterol. SI conversion factor: To convert cholesterol to millimoles per liter, multiply by *Unless otherwise indicated, data are expressed as percentage of subjects. Percentages have been rounded and might not total 100. Indicates the use of digoxin, nitrates, or antiarrhythmic drugs. (n=871) or persons with insufficient data to exclude previous myocardial infarction (n=186); (2) participants with angina pectoris as measured with the Rose questionnaire (n=454) or participants who had not completed the Rose questionnaire (n=103); (3) participants with a history of coronary artery bypass grafting or percutaneous transluminal coronary angioplasty (n=195); and (4) participants with atrial fibrillation at baseline (n=376). Thus, subjects for whom an ECG was obtained at baseline and an ultrasonographic examination had been performed (n=4407) were included in the present study. Among these, 182 participants had missing information on the intima-media thickness measurement and 215 had missing information on carotid plaques. Measurements of the covariates in the population for analysis were complete for more than 99% of the study population, except for hypertension and body mass index (1.5% and 2.0% missing data, respectively). Participants with missing data on covariates were excluded in the analyses. STATISTICAL ANALYSES The distribution of the carotid intima-media thickness measurements was positively skewed. We applied a natural logarithmic transformation to obtain a more normal distribution. The natural log-transformed intima-media thickness was categorized in quartiles based on the distribution of the population for analysis. Participants with a carotid plaque score of 0, Table 2. Measures of Atherosclerosis and Risk of Atrial Fibrillation: the Rotterdam Study, No. of RR (95% CI) Characteristic Cases/ Subjects Model 1* Model 2 Carotid intima-media thickness, quartile First 30/ [Reference] 1 [Reference] Second 42/ ( ) 1.13 ( ) Third 86/ ( ) 1.99 ( ) Fourth 98/ ( ) 1.90 ( ) None 80/ [Reference] 1 [Reference] Mild 47/ ( ) 1.33 ( ) Moderate 46/ ( ) 1.19 ( ) Severe 77/ ( ) 1.49 ( ) Abbreviations: CI, confidence interval; RR, relative risk. *Adjusted for age and sex. Adjusted for age, sex, body mass index, hypertension, systolic blood pressure, total serum cholesterol level, smoking, diabetes mellitus, left ventricular hypertrophy on the electrocardiogram, and cardiac medication (use of digoxin, nitrates, or antiarrhythmic drugs). 1, 2, and 3 or more were categorized as having no, mild, moderate, and severe carotid atherosclerosis, respectively. In the analyses, the lowest carotid intima-media thickness quartile and the absence of carotid plaques were used as the reference categories. Age- and sex-adjusted hazard rate ratios, expressed as relative risks with their 95% confidence intervals, were calculated using the Cox proportional hazards model with atrial fibrillation as the dependent variable and the atherosclerosis variables as the independent variables. Additional adjustments were made for body mass index, hypertension, systolic blood pressure, serum cholesterol level, smoking, diabetes mellitus, left ventricular hypertrophy on the ECG, and the use of cardiac medication (digoxin, nitrates, and antiarrhythmic drugs). The analyses were further stratified by sex. In a secondary analysis, we examined the associations after exclusion of those subjects in whom incident myocardial infarction had preceded the onset of atrial fibrillation during the follow-up period. The data were managed and analyzed using SPSS for Windows statistical software (release 11.0; SPSS Inc, Chicago, Ill). Population-attributable risks with 95% confidence intervals, adjusted for confounders, were calculated with IRAP, version ,20 RESULTS The baseline characteristics of the study population are summarized in Table 1. During a median follow-up of 7.5 years, 269 new cases (6.1%) of atrial fibrillation were identified. Of those, 44 cases (16.4%) were identified by the ECG obtained at the research center only (n=38) or were identified by ECG earlier than by the 2 other ascertainment methods (n=6). After adjustments for age and sex, statistically significant associations were found between carotid intimamedia thickness and carotid plaques and the risk of atrial fibrillation. The associations were stronger for the higher quartiles of intima-media thickness and increased severity of carotid plaques. After further adjustments for clinical features, the associations were somewhat attenuated but remained statistically significant (Table 2). 384

4 Separate results for men and women are presented in Table 3. In women, the associations between the atherosclerosis measures and risk of atrial fibrillation were positive and statistically significant and reflected a doseresponse relationship. In men, however, the association between intima-media thickness and risk of atrial fibrillation increased with increasing quartiles of intimamedia thickness, but this association was not statistically significant. No association was found between carotid plaques and risk of atrial fibrillation in men. When we repeated the analyses excluding those new atrial fibrillation cases that developed after an incident myocardial infarction (n=13), the associations were not materially different (data not shown). The population-attributable risks associated with intima-media thickness and carotid plaques were 33% and 19%, respectively. COMMENT In this population-based study, we found that atherosclerosis was associated with new-onset atrial fibrillation in subjects without a history of coronary heart disease. The indices of atherosclerosis used in this study have been shown to be measures of generalized atherosclerosis. 21,22 The results suggest that atherosclerosis may be a causal factor in the etiology of atrial fibrillation, even when no overt coronary heart disease is present. To our knowledge, the Cardiovascular Health Study is the only population-based study that previously investigated associations between measures of atherosclerosis and atrial fibrillation. 6,7 In that study, carotid intimamedia thickness and the ankle-arm index were not associated with incident atrial fibrillation. The reason for the discrepancy with our findings is not clear. There are, however, important differences in populations, ascertainment methods, and analyses between the 2 studies. For example, the Cardiovascular Health Study included self-reported cases of atrial fibrillation, and the proportion of those who had a history of a myocardial infarction in the Cardiovascular Health Study was lower than that in the Rotterdam Study (9.2% vs 13.2%), despite the older mean age of the Cardiovascular Health Study population. 23 The older age of the Cardiovascular Health Study population is notable, because one could speculate that associations between atherosclerosis variables and atrial fibrillation may be reduced at an older age through competing vascular risks. In addition, associations in the Cardiovascular Health Study were adjusted for left atrial size, valvular disease, and a coronary injury score based on ECG characteristics that may be highly correlated with atherosclerosis and thus could obscure a relation between atherosclerosis and atrial fibrillation. An analysis from the Manitoba Follow-up Study, 8 in which angina without myocardial infarction and ST-T wave changes without clinical evidence of ischemic heart disease were independently associated with the onset of atrial fibrillation, supports our hypothesis that atherosclerotic vascular disease is associated with atrial fibrillation. Associations of carotid intima-media thickness and risk of atrial fibrillation were weaker in men than in women, and the association between carotid plaques and atrial Table 3. Measures of Atherosclerosis and Risk of Atrial Fibrillation by Sex: the Rotterdam Study, RR (95% CI) Characteristic Cases/ Subjects Model 1* Model 2 Men Carotid intima-media thickness, quartile First 12/317 1 [Reference] 1 [Reference] Second 15/ ( ) 1.01 ( ) Third 30/ ( ) 1.67 ( ) Fourth 40/ ( ) 1.61 ( ) None 34/637 1 [Reference] 1 [Reference] Mild 18/ ( ) 1.18 ( ) Moderate 10/ ( ) 0.55 ( ) Severe 32/ ( ) 1.12 ( ) Women Carotid intima-media thickness, quartile First 18/739 1 [Reference] 1 [Reference] Second 27/ ( ) 1.20 ( ) Third 56/ ( ) 2.26 ( ) Fourth 77/ ( ) 2.14 ( ) None 46/ [Reference] 1 [Reference] Mild 29/ ( ) 1.46 ( ) Moderate 36/ ( ) 1.73 ( ) Severe 45/ ( ) 1.84 ( ) Abbreviations: CI, confidence interval; RR, relative risk. *Adjusted for age and sex. Adjusted for age, sex, body mass index, hypertension, systolic blood pressure, total serum cholesterol level, smoking, diabetes mellitus, left ventricular hypertrophy on the electrocardiogram, and cardiac medication (use of digoxin, nitrates, or antiarrhythmic drugs). fibrillation was absent in men. We are unable to fully explain these differences between the sexes, but one possibility is that men with a high degree of atherosclerosis have died of cardiovascular disease before entering our study or before the development of atrial fibrillation. Reliable associations strongly depend on accurate incidence figures. Population-based incidence figures of atrial fibrillation, however, are rare. Our incidence rates are comparable to those reported by the Framingham Study. The incidence figures of the Cardiovascular Health Study, however, are twice as high and probably reflect the older age of that cohort and other ascertainment methods. We could hypothesize that atherosclerosis, gradually or abruptly, reduces the blood supply to the sinus node and the atrial tissues that affect the gradual spreading of the electrical impulse over the atria and a normal contraction of the atria. This reduced flow may cause fibrosis and microscopic scarring of the atrial wall, 2 resulting in areas with reduced conduction or even blocked conduction. Areas in the atria with decreased conduction velocity have been shown to favor reentry mechanisms, which can result in the development of atrial fibrillation. 24,25 Ischemia itself, however, may also shorten the refractory period and decrease the conduction velocity, potentially facilitating reentry processes and the onset of atrial fibrillation. 26 These hypotheses are also supported by several pathology and experimental studies. 3,4,27 For 385

5 example, in a study of subjects with longstanding atrial fibrillation, 4 nodal artery stenosis was present without distinct lesions of the sinus node or the internodal tracts, suggesting that diminished blood flow may play a role in the origin of atrial fibrillation in these cases. In another large pathological study, 3 approximately 16% of subjects with longstanding atrial fibrillation had moderate to severe coronary artery disease without any histological evidence of infarction. Finally, it has been postulated that the sinus node artery plays an essential role in the synchronization of the numerous sites of automaticity within the sinus node. Dysfunction or rigidity of the sinus node artery could thus lead to a state of decreased rhythmic stability. 28 Our study is limited by the assumption that measures of noncoronary atherosclerosis are markers of the extent of coronary atherosclerosis. However, a previous investigation from our group 22 found that the measures of noncoronary atherosclerosis, including carotid intima-media thickness and carotid plaques, were strongly associated with coronary calcification, as measured by electron beam computed tomography. Not all participants in the Rotterdam Study who completed the baseline examinations had an ultrasonographic examination, owing to a restricted availability of ultrasonographers at the end of 1992 and in Therefore, missing data on the ultrasonographic examinations were random and not based on the disease status of the participants. We did not distinguish between atrial fibrillation (paroxysmal or permanent) and atrial flutter when identifying cases, but these arrhythmias have the same risk factors and the same consequences. 29,30 Moreover, periods of atrial fibrillation in paroxysmal atrial fibrillation are much more common than is generally perceived by patients or their physicians and, over time, a high proportion of paroxysmal atrial fibrillation changes into chronic atrial fibrillation. 31,32 In aging populations, as is the case in Western Europe, atrial fibrillation will become progressively more important as its prevalence increases sharply with age. 13 Atrial fibrillation is associated with considerable morbidity and mortality, increasing numbers of hospital admissions, and rising costs Our findings suggest that the effect of atherosclerosis on the development of atrial fibrillation may be larger than is commonly thought. The results indicate that the management of atherosclerotic vascular disease and its risk factors in the general population and the aggressive treatment of participants with higher levels of atherosclerosis may contribute to a decrease in the incidence of atrial fibrillation, thereby lowering the disease burden and costs. The population-attributable risks for intima-media thickness of 33% and for carotid plaques of 19% suggest that successful preventive treatment of atherosclerosis may result in a considerable reduction of atrial fibrillation cases in the community. In conclusion, in this prospective, population-based study, we found that measures of generalized atherosclerosis predict the onset of atrial fibrillation in persons without obvious coronary heart disease. The findings suggest that aggressive management of atherosclerotic vascular disease, including in asymptomatic subjects, may help to prevent atrial fibrillation. Accepted for Publication: October 30, Correspondence: Jacqueline C. M. Witteman, PhD, Department of Epidemiology and Biostatistics, Erasmus Medical Center, PO Box 2040, 3000 CA Rotterdam, the Netherlands (j.witteman@erasmusmc.nl). Author Contributions: Study concept and design: Heeringa, Hofman, and Witteman. Acquisition of data: Heeringa, Kors, and van Rooij. Analysis and interpretation of data: Heeringa, van der Kuip, Lip, and Witteman. Drafting of the manuscript: Heeringa. Critical revision of the manuscript for important intellectual content: Heeringa, van der Kuip, Hofman, Kors, van Rooij, Lip, and Witteman. Statistical analysis: Heeringa and Witteman. Administrative, technical, and material support: Heeringa, van Rooij, and Lip. Study supervision: Hofman, Kors, and Witteman. Financial Disclosure: None reported. REFERENCES 1. Benjamin EJ, Levy D, Vaziri SM, D Agostino RB, Belanger AJ, Wolf PA. Independent risk factors for atrial fibrillation in a population-based cohort: the Framingham Heart Study. JAMA. 1994;271: Anderson KR, Sutton MG, Lie JT. Histopathological types of cardiac fibrosis in myocardial disease. J Pathol. 1979;128: Aberg H. Atrial fibrillation, I: a study of atrial thrombosis and systemic embolism in a necropsy material. Acta Med Scand. 1969;185: Davies MJ, Pomerance A. Pathology of atrial fibrillation in man. Br Heart J. 1972; 34: Falk RH. Etiology and complications of atrial fibrillation: insights from pathology studies. Am J Cardiol. 1998;82:10N-17N. 6. Furberg CD, Psaty BM, Manolio TA, Gardin JM, Smith VE, Rautaharju PM. Prevalence of atrial fibrillation in elderly subjects (the Cardiovascular Health Study). Am J Cardiol. 1994;74: Psaty BM, Manolio TA, Kuller LH, et al. Incidence of and risk factors for atrial fibrillation in older adults. Circulation. 1997;96: Krahn AD, Manfreda J, Tate RB, Mathewson FA, Cuddy TE. The natural history of atrial fibrillation: incidence, risk factors, and prognosis in the Manitoba Follow-up Study. Am J Med. 1995;98: Kannel WB, Abbott RD, Savage DD, McNamara PM. Coronary heart disease and atrial fibrillation: the Framingham Study. Am Heart J. 1983;106: Cameron A, Schwartz MJ, Kronmal RA, Kosinski AS. Prevalence and significance of atrial fibrillation in coronary artery disease (CASS Registry). Am J Cardiol. 1988;61: Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE. Common carotid intima-media thickness and risk of stroke and myocardial infarction: the Rotterdam Study. Circulation. 1997;96: Hofman A, Grobbee DE, de Jong PT, van den Ouweland FA. Determinants of disease and disability in the elderly: the Rotterdam Elderly Study. Eur J Epidemiol. 1991;7: Heeringa J, van der Kuip DA, Hofman A, et al. Prevalence, incidence and lifetime risk of atrial fibrillation: the Rotterdam study. Eur Heart J. 2006;27: van Bemmel JH, Kors JA, van Herpen G. Methodology of the modular ECG analysis system MEANS. Methods Inf Med. 1990;29: Willems JL, Abreu-Lima C, Arnaud P, et al. The diagnostic performance of computer programs for the interpretation of electrocardiograms. N Engl J Med. 1991; 325: Kors JA, van Herpen G, Wu J, Zhang Z, Prineas RJ, van Bemmel JH. Validation of a new computer program for Minnesota coding. J Electrocardiol. 1996;29: Rose GA. The diagnosis of ischaemic heart pain and intermittent claudication in field surveys. Bull World Health Organ. 1962;27: Accessed October van der Meer IM, Bots ML, Hofman A, del Sol AI, van der Kuip DA, Witteman JCM. Predictive value of noninvasive measures of atherosclerosis for incident myocardial infarction: the Rotterdam Study. Circulation. 2004;109: Engel LS, Chow WH, Vaughan TL, et al. Population attributable risks of esophageal and gastric cancers. J Natl Cancer Inst. 2003;95: National Cancer Institute. Interactive Risk Attributable Program (IRAP v2.2). http: //dceg.cancer.gov/irap/. Accessed October

6 21. Rohani M, Jogestrand T, Ekberg M, et al. Interrelation between the extent of atherosclerosis in the thoracic aorta, carotid intima-media thickness and the extent of coronary artery disease. Atherosclerosis. 2005;179: Oei HH, Vliegenthart R, Hak AE, et al. The association between coronary calcification assessed by electron beam computed tomography and measures of extracoronary atherosclerosis: the Rotterdam Coronary Calcification Study. JAm Coll Cardiol. 2002;39: Psaty BM, Kuller LH, Bild D, et al. Methods of assessing prevalent cardiovascular disease in the Cardiovascular Health Study. Ann Epidemiol. 1995; 5: Moe GK, Abildskov JA. Atrial fibrillation as a self-sustaining arrhythmia independent of focal discharge. Am Heart J. 1959;58: Cox JL, Canavan TE, Schuessler RB, Boineau JP. The surgical treatment of atrial fibrillation, II: intraoperative electrophysiologic mapping and description of the electrophysiologic basis of atrial flutter and atrial fibrillation. J Thorac Cardiovasc Surg. 1991;101: Cranefield PF, Wit AL, Hoffman BF. Genesis of cardiac arrhythmias. Circulation. 1973;47: James TN, Hershey EA. Experimental studies on the pathogenesis of atrial arrhythmias in myocardial infarctions. Am Heart J. 1962;63: James TN. Structure and function of the sinus node, AV node and his bundle of the human heart, II: function. Prog Cardiovasc Dis. 2003;45: Halligan SC, Gersh BJ, Brown RD Jr, et al. The natural history of lone atrial flutter. Ann Intern Med. 2004;140: Lelorier P, Humphries KH, Krahn A, et al. Prognostic differences between atrial fibrillation and atrial flutter. Am J Cardiol. 2004;93: Israel CW, Gronefeld G, Ehrlich JR, Li YG, Hohnloser SH. Long-term risk of recurrent atrial fibrillation as documented by an implantable monitoring device: implications for optimal patient care. J Am Coll Cardiol. 2004;43: Kato T, Yamashita T, Sagara K, Iinuma H, Fu LT. Progressive nature of paroxysmal atrial fibrillation: observations from a 14-year follow-up study. Circ J. 2004; 68: Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22: Kannel WB, Abbott RD, Savage DD, McNamara PM. Epidemiologic features of chronic atrial fibrillation: the Framingham study. N Engl J Med. 1982;306: Benjamin EJ, Wolf PA, D Agostino RB, Silbershatz H, Kannel WB, Levy D. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation. 1998;98: Stewart S, MacIntyre K, MacLeod MM, Bailey AE, Capewell S, McMurray JJ. Trends in hospital activity, morbidity and case fatality related to atrial fibrillation in Scotland, Eur Heart J. 2001;22: Stewart S, Murphy N, Walker A, McGuire A, McMurray JJ. Cost of an emerging epidemic: an economic analysis of atrial fibrillation in the UK. Heart. 2004; 90: Boriani G, Diemberger I, Martignani C, Biffi M, Branzi A. The epidemiological burden of atrial fibrillation: a challenge for clinicians and health care systems. Eur Heart J. 2006;27: